Reprints Available Directly from the Publisher Photocopying Permitted by License Only a Selective Culture System for Generating Terminal Deoxynucleotidyl Transferase-positive Lymphoid Cells in Vitro. v. Detection of Stage-specific Pro-b-cell Stimulating Activity in Medium Conditioned by Mouse Bone M

The selective in vitro generation of rat, mouse, and human terminal deoxynucleotidyl transferase-positive (TdT lymphoid cells in our long-term xenogeneic bone marrow (BM) culture system is characterized by physical interaction between the developing lymphocytes and mouse BM-adherent stromal cells and macrophages. In the present study, experiments in which micropor)us membrane culture inserts were inoculated with rat BM cells demonstrated that although the generation of primitive B-lineage lymphoid cells requires the presence of a mouse BM feeder layer, cognitive recognition events are not necessary. Similarly, cell-free (and serum-free) medium conditioned with mouse BM (but not thymus or spleen) adherent cells and stromal-cell lines therefrom supported the proliferation of early rat lymphoid cells in a dose-dependent manner. Double immunofluorescence for incorporated bromo-deoxyuridine (BrdU) and early B-lineage markers of rat BM lymphoid cells maintained in culture inserts or conditioned medium (CM), and studies of their in vitro and in vivo developmental potentials, indicated that the lymphoproliferative response resulted from the selective stimulation of lymphoid stem and/or progenitor cells. The most primitive of these target cells had a HIS24 HIS50-TdT-c/-sIg-, pre-pro-B-cell phenotype. Whereas this subset normally constitutes less than 2% of B-lineage BM cells in vivo, it comprises more than 25% of total lymphoid cells in vitro. In addition, the number of TdT cells, predominantly of the early pro-B-cell phenotype (HIS24 HIS50-c/-sIg-), was increased approximately tenfold above input levels. Based on these and previous findings, a schematic model is proposed for the developmental pathway of early B-lineage cells in rat BM from the level of the committed (possibly common) lymphoid stem cell to that of the pre-B-cell.


in vitro bone
marrow (BM) culture systems, in which manipula- "Corresponding author.tion of culture conditions profoundly affects the lineages and/or stages of development that are generated.For example, Dexter-type cultures selec- tively produce myeloid lineage cells and multipoten- tial stem cells, but few lymphoid cells (Dexter et al.,  1977), whereas the Whitlock-Witte culture system selectively produces large numbers of pre-B cells (c/ +, sIg-), but not myeloid or multipotent stem cells (Whitlock and Witte, 1982; Whitlock et al.,  1984).Recently, modifications to the Whitlock- Witte culture system have been described that permit the generation of B-lineage cells more primitive than those observed in standard cultures.These modifications include seeding the BM feeder layers with mouse fetal liver cells (Denis et al., 1984,   1987), transfecting the cultured ells with the BCR/ ABL chimeric oncogene (Scherele et al., 1990), and initiating the culture in the presence of interleukin-4 (Peschel et al., 1989).

The role of oluble factors in the generation of pre-B-cells has been extensively studied using bone marrow stromal cell lines established from the foregoing cultures.In particular, interleukin-7 (IL-7), a cytokine first purified from mouse BM stromal cells, has been found to stimulate the proliferation of pre-B (B220/) and possibly pro-B (B220-) cells from mouse BM (Namen et al., 1988a,1988b; Henney,   1989; Lee et al., 1989).However, despi e the effect of IL-7, the long-term survival of early B-lineage cells in primary cultures generally is not maintained in the absence of direct contact with BM stromal cells ( Kierney and Dorshkind, 1987; Sudo et al.,   1989).This is consistent with the development of foci of proliferating lymphoid precursor cells on and within the adherent BM stromal-cell layer (Whitlock  and Witte, 1982; Whitlock et al., 1984; Hayashi et  al., 1984; Medlock et al., 1993a, 1993b).Hardy et al. (1991) have reported that the earliest phenotypically distinct mouse B-lineage cell population, called "pre-pro-" B-cells (B220 $7 / BP-I-, HSA-, Ig genes in germline configuration), requires only contact with BM stromal cells to survive in a 4-day coculture system.However, later stages of lymphopoiesis, marked by expression of HSA and progressive Ig gene rearrangements, were increas- ingly feeder-layer adherence-independent and IL-7- dependent (Nishikawa et al., 1988; Sudo et al.,  1989; Hardy et al., 1991).These observations sug- gest that close-range molecular interactions between lymphoid progenitor cells and microenvironmental cells regulate the initial stages of lymphopoiesis in BM.Although the nature of the requirement for close lymphoid-precursor:stromal-cell association is not known, a potential candidate for a stromal cell-bound second signal is the recently described stem-cell factor (SCF), which can act synergistically with IL-7 to stimulate lymphopoiesis (McNiece et  al., 1991; Billips et al., 1992).

To specifically study the regulatory mechanisms that are operational at the most primitive stages of lymphoid development, we have designed a long- term xenogeneic BM culture system that selectively supports the generation of normal (and leukemic)   lymphoid precursor cells from adult rat, mouse, and human BM on a mouse BM-adherent cell-feeder layer (Hayashi et al., 1984; Medlock et al., 1987a;   Goldschneider and King, 1991; Medlock et al.,   1993a, 1 93b).The least mature lymphoid cells in the cultures of rat BM have a HIS24 / HIS50- antigenic phenotype (B220 / HSA-equivalent), lack the enzyme terminal deoxynucleotidyl transferase (TdT), and are mostly adherent to the mouse BM feeder layer.In contrast, the most mature cells in these cultures are HIS24 HIS50 c#-late pro-Bcells, and are present primarily in the nonadherent phase of the culture.

Despite these observations, it is not clear that adherence to the feeder layers is essential for the maintenance of lymphopoiesis in our culture system.Thus, although lymphoid precursor cell activity is approximately tenfold higher per unit number of cells in the adherent than in the nonadherent phase of the culture (Medlock et al., 1993b), approximately 20% of the nonadherent lymphoid cells are in DNA synthesis (Hayashi et al., 1984).Moreover, reduc- tio of the concentration of fetal bovine serum (FBS) in the culture medium markedly decreases the per- centage of adherent TdT / lymphoid cells, but not the total number of lymphoid cells generated (Medlock et al., 1993b).Therefore, the requirement for physical contact between the developing lymphoid cells and the mouse BM-adherent cells was formally investigated in the present study.The results dem- onstrate that rat BMTdT / pro-B-cells and their TdT-precursors (pre-pro-B-cells) can be generated both in microporous membrane culture inserts placed over mouse BM-adherent cell-feeder layers, as well as in serum-free medium conditioned by stromal cells from these feeder layers.Hence, al- though contact between these primitive lymphoid precursors and mouse BM stromal cells may opti- mize lymphopoiesis in our culture system, the generation of pro-B-cells is, in the final analysis, maintained by a soluble factor(s) from such stromal cells.


RESULTS

Selective Generation of Rat BM Lymphoid Precursor Cells in Culture Inserts To examine whether physical contact between the mouse BM-adherent feeder-layer cells and the rat BM lymphoid precursor cells is required for the generation of TdT / lymphoid cells in vitro, freshly harvested rat BM cells were cultured in microporous membrane culture inserts placed over (but not in contact with) confluent mouse BM feeder layers for 10 days.As demonst

ted in F
g. 1A, rat BM lymphoid cells were selectively maintained in these culture inserts, but not in inserts placed in wells lacking a feeder layer.Typically, lymphoid cells accounted for 60-90% of total cells recovered from the culture inserts.Of these, approximately 30-50% were TdT/, as compared with 3-5% of total cells in the starting inoculum.

Results in Fig. 1B show that lymphopoiesis could be maintained for at least 3 weeks by serially pas- saging the culture inserts onto fresh mouse BM feeder layers at 10-day intervals.However, when the culture inserts were transferred to wells contain- ing rat (instead of mouse) BM feeder layers, or culture medium only, the lymphoid cells rapidly died.This is consistent with our earlier observation that rat BM feeder layers do not, by themselves, support BM lymphopoiesis in vitro, even when direct ell contact is permitted (Hayashi et al., 1984).

As in the standard culture system, only extremely immature B-lineage cells, almost all of which ex- press the HIS24 marker (Opstelten et al., 1986), were maintained in the culture inserts (CI).Thus, as compared with the original BM-cell inoculum, day 10 CI-generated lymphoid cells were completely depleted of sIg / B cells and c/ + pre-B-cells, and partially depleted (approximately twofold) of inter- mediate and late pro-B-cells (HIS50 / TdT / and HIS50 / TdT-, respectively) (Table 1).Conversely the CI-generated lymphoid cells were enriched ap- proximately twofold for the HIS50-TdT-subset of pre-pro-B-cells, and tenfold for the HIS50-TdT / subset of early pro-B-cells, which together consti- tuted 70% of the total lymphoid cells present.The developmental relationships of these phenotypic subsets have been established in previous experiments in which HIS24 / HIS50-TdT-cells generated HIS24 / HIS50-TdT / and thence HIS24 / HIS50 / TdT / and HIS24 / HIS50 / TdT-cells in vitro (Goldschneider et al., 1987; Hunt et al., 1988;  Medlock et al., 1993b).Similarly, HIS24 / HIS50- cells in normal rat BM are thought to be the immediate precursors of HIS24 HIS50 / pro-B-cells in vivo (Hermans, 1991).

Although the proportions of rat TdT / and TdT- lymphoid cells generated in culture inserts was similar to that generated directly on mouse BM- adherent cell-feeder layers, the number of total lymphoid cells present was approximately tenfold lower.One possible explanation is that direct con- tact of the lymphoid precursor (or other) cells in rat BM with the mouse BM feeder-layer cells may increase the production and/or release of lympho- stimulatory factors into the medium (Sudo et al., 1989).To address this possibility, rat BM cells were cultured for 10 days in microporous membrane in- serts placed over mouse BM feeder layers that had also been seeded with rat BM cells.Although nor- mal numbers of lymphoid cells were generated on the feeder layers, the number of lymphoid cells recovered from the culture inserts placed over the seeded feeder layers was approximately 25% lower (rather than higher) than that from inserts placed over unseeded feeder layers (data not shown).


A

Mouse BM


B None

Mouse BM Rat BM Adherent Cell Feeder Layer None FIGURE 1. Maintenance of rat BM lymphoid precursor cells in culture inserts. 5 10 freshly harvested rat BM cells were added to microporous membrane culture inserts placed in wells in the presence or absence of a mouse BM-adherent cell-feeder layer.

Results represent the mean number (+S.D.) of total (solid) and TdT (hatched) lymphoid cells per culture insert:

(A) 10 days later, and (B) 10 days after transfer of the ul- ture inserts to new wells in the presence or absence of a mouse or rat BM-adherent cell-feeder layer (20 days total elapsed time in vitro).Similar numbers of TdT lymphoid cells were maintained in culture in- ser s upon transfer to tertiary feeder layers (30 days total elapsed time).Pre-pro-B


Stage of development


Subset

TdT-HIS50- 0.8 1.5 + 1.9-fold aFreshly harvested rat BIv cells placed in culture inserts (5 X 10 cells/insert) confluent BM-adherent cell layers day 0. Ten days later, cells recovered for immunofluorecence analysis.

Another possible explanation for the greater generation of lymphoid cells in the standard culture system than in culture inserts is that some of the lymphostimulatory activity normally is bound to components of the extracellular matrix (ECM) and/ or cell membranes in the feeder layer (Gordon et al.,  1987; Roberts et al., 1988).To test this, mouse BM-adherent cell-feeder layers were extracted with 2M NaC1 solution (Gordon et al., 1987), desalted by ultra iltration (MW cutoff 10 kD) in serum-free cul- ture medium, and brought up to 25 % FBS.Although the saline solution did not affect the lymphostimulatory activity in medium conditioned by mouse BM feeder layers (see what follows), no lymphostimu- latory activity was detected in the saline extract from these feeders (data not shown).

Selective Generation of Rat BM TdT + Lymphoid Cells in Conditioned Medium (CM) To more directly examine the role of soluble factors in the generation of primitive lymphoid cells in vitro, freshly harvested rat BM cells were cultured in medium that had been conditioned for 10 days with mouse BM-adherent cell-feeder layers.After 8 days in the CM, there was a readily discernible increase in the percentage of rat TdT + lymphoid cells over that in control medium.However, it was difficult to quantify this increase due to the presence of large numbers of dead myeloid cells.To circumvent this, enriched suspensions of rat BM lymphoid cells, generated in standard 10-day cultures, were substituted for freshly harvested rat BM cells in the CM.As shown in Figs. 2 and 3, the lymphoid cells were maintained for at least 4 days (>90% viability) by CM from mouse BM feeder layers, but not by CM from mouse thymic or splenic adherent cells or rat BM feeder layers.Moreover, approximately 50% of the accumulated lymphoid cells expressed TdT.

Results in Table 2 demonstrate that the rat TdT + BM lymphoid cells recovered from CM after 4 days were phenotypically similar to those harvested from culture inserts and standard cultures, supporting the notion that pro-B-cells in these cultures are selec- tively stimulated by soluble mediators from the mouse BM-adherent cell feeder.This was further tested in experiments in which culture-generated rat BM lymphoid cells, incubated for 4 days in CM, were pulsed with BrdU.As shown in Table 3, approximately one-third of the lymphoid cells in- corporated BrdU and all of these had the HIS24 + HIS50-phenotype.Conversely, approxi-


Mouse BM Mouse Mouse

Rat BM None

Thymus Spleen Source of Conditioned

edium FIGUR
as conditioned for 10 days by adherent cells from the indicated tissues and diluted twofold in nonconditioned medium.

5 105 culture-generated rat BM lymphoid cells were incubated for 4 days in 2 ml of CM. Results represent the mean number (+S.D.) of total (solid) and TdT (hatched) lymphoid cells per well.

FIGURE 3. Morphology of rat BM cells generated in the absence of direct contact with mouse BM feeder layers.The lymphoid cells were originally generated for 10 days in microporous membrane culture inserts (A, B) or in standard cultures (C), and the lymphoid cells were cultured for an additional 4 days in medium conditioned by mouse BM-adherent cells (A, C) or in normal medium (B).

Numerous immature lymphoid cells, some undergoing mitosis (arrow), and occasional macrophages are present in (A) and (C), whereas only macrophages are present in (B).May Grunwald-Giemsa stain. 1000.Mouse MHC (H-2b)
< < < < Surface Ig < < < < Cytoplasmic/ < 1 < 1 < < 1 Ox-19 (Pan T cell) < < 1 < < Ox-39 (IL-2R) < < < < HIS24 (CD45R-B220) > 95 > 95 > 95 > 95
HIS50 (HSA) 35 50 49 96 aCells analyzed by double immunofluorescence for TdT and indi ated markers (1)culture-generated rat BM lymphoid precursor cells incubated in CM, (2)rat BM lymphoid precursor cells generated in culture inserts in the presence of BM-adherent cell-feeder layer, (3) nonadherent rat lymphoid cells recovered from standard cultures, and (4)freshly harvested rat BM cells.mately two-thirds of the HIS24 / HIS50-cells incor- porated BrdU, and approximately 50% of these were TdT / It should be noted that incubation with BrdU did not alter the total number of lymphoid cells recovered from the CM-treated cultures.


Dose-response experiments demonstrate that

both the number of lymphoid cells recovered from the CM (Fig. 4A) and the incorporation of tritiated thymidine by these cells (Fig. 4B) was proportional to the concentration of CM after ultrafiltration, as was the frequency of mitotic figures in cell smears (Fig. 3).In contrast, concentrated unconditioned me- dium had no det ctible lymphostimulatory effect.Furthermore, medium conditioned with mouse BM feeder layers that had been seeded with rat BM cells had less lymphostimulatory activity than did CM from unseeded mouse BM feeder layers (data not shown).

Because the culture-generated rat BM lymphoid cells in the preceding experiments might have been contaminated with mouse BM cells from the feeder layer, the ability of CM to support the growth of rat BM lymphoid cells, generated exclusively in culture inserts, was also tested.The results indicated that rat BM lymphoid cells' from both sources were main- tained equally well by CM (Fig. 3 and data not shown).This strongly suggests that the CM stimu- lates early lymphopoiesis in rat BM by acting di- rectly on the lymphoid precursors.

We next determined whether active CM could be generated in the absence of serum.Mouse BM feeder layers were intially cultured for 10 days in medium containing 25% FBS, after which they were washed extensively with and cultured for an additional 4 days in serum-free RPMI-1640.(The feeder layers could be maintained for 4 to 5 days in this serum- free medium before showing signs of deterioration.)The serum-free CM (SFCM) was then concentrated up to twentyfold by ultrafiltration and added to culture-generated rat BM lymphoid precursor cells in the presence of 25 % FBS.As demonstrated in Fig. 5, the number of lymphoid cells obtained 4 days later was a function of the concentration of SFCM.It was further observed that lymphostimulatory activity comparable to that in ten-fold concentrated SFCM was obtained in unconcentrated day 10 CM gener-  CM was concentrated tenfold by ultrafiltration and diluted with nonconditioned medium to the indicated final concentrations,

where is equivalent to unconcentrated, undiluted CM.Units are given with respect to unconcentrated CM. (A)5x105 culture- generated rat BM lymphoid cells were incubated for 4 days in 2 ml of CM. Results represent the mean (+S.D.) number of total (solid line) and TdT (dashed line) lymphoid cells per well.(B) 105 culture-generated rat BM lymphid cells were placed in 0.2 ml CM for 72 hr.Twelve hours prior to harvesting, wells were pulsed with mCi[3H]-TdR.Results represent c.p.m, of [3H]-TdR incorporated by total lymphoid cells per well in a representative experiment (one of three).

ated in medium containing 1% FBS and 5% Nutri- doma (data not shown).

Developmental Potential of CM-Sensitive Rat Lymphoid Cells As the lymphoid cells that were stimulated by CM appeared to have a very immat

e phenotype, it was of interest to determine whether any of these cells could func
ion as lymphoid progenitors in vitro when replated in standard cultures.The results in Fig. 6 show that after incubating freshly harvested rat BM cells in CM for 4 days or culture inserts for " 10 2.5 5 10 20

Concentration of Serum-Free Conditioned Medium FIGURE 5. Maintenance of rat BM lymphoid precursor cells in various concentrations of serum-free conditioned medium.The SFCM, prepared as described in Methods, was collected from mouse BM-adherent cells after 4 days, concentrated twentyfold by ultrafiltration and diluted with serum-free medium to the indicated concentration, where 1 is equivalent to unconcentrated, undiluted SFCM. 5 105culture-generated rat BM lymphoid pre- cursor cells were cultured for 4 days in 2 ml of SFCM after being supplemented with FBS to a final concentration of 25%.Results represent the number of total (solid line) and TdT (dashed line) lymphoid cells per well.

10 days, progenitor activity approximately equiva- lent to that present in the nonadherent phase of standard cultures was recovered.These progenitor cells, after transfer to the feeder layers, formed lymphoproliferative foci on and within the mouse BM-adherent cells.Conversely, only minimal pro- genitor activity was detected in rat BM cells follow- ing 4 days of culture in control medium.

To test the in vivo lymphopoietic potential of the CM-sensitive BM cells, freshly harvested rat BM cells were cultured in CM for 4 days prior to being injected intravenously into sublethally irradiated, RT-7 and Igk-1 alloantigen disparate rats.After 25 days, the spleens and thymuses of the recipient rats were collected and stained with antibodies specific for donor B cells (anti-Igklb) and T cell (anti-RT-7b).Unlike human and mouse B cells, approximately 95% of rat B cells express the a:Ig light-chain isotype (Springer et al., 1982), so that it is a useful allogspecific marker for sIg+ B cells.As demonstrated in Fig. 7, rat BM cells precultured in CM, as those obtained from standard cultures, maintained a significantly igher capacity to regenerate both the B-and T-cell compartments in the recipient rats than those cells that were precultured in control medium.Conversely, the level of pluri-  freshly harvested rat BM cells were incubated for 4 days in medium conditioned with mouse BM adherent cells or in culture inserts placed over mouse BM-adherent cell-feeder layers for 10 days.The cells were then transferred directly onto confluent mouse BM feeder layers for an additional 10 days.Results

represent the mean number (+S.D.) of total (solid) and TdT (hatched) rat lymphoid cells recovered in the nonadherent phase of these secondary cultures, as compared with the number of lymphoid cells generated 10 days after seeding X10 freshly harvested rat BM cells directly onto mouse BM feeder layers.

potent stem-cell activity, as detected indirectly by CFU-S assay, decreased more than 90% during culture in both CM and control medium, and more than 95% in standard cultures, as shown previously (Hayashi et al., 1984).

Constitutive Release of Lymphostimulatory Activity by Mouse BM Stromal-Cell Lines Mouse BM-adherent cell-feeder layers were repeat- edly passaged at 10-day intervals in order to gener- ate stromal-cell lines.After approximately 2 months, two lines of adherent cell (F12-5B6 and F1-12B6), each having a homogenous stromal-cell morphology, were isolated.As shown in Fig. 8, these cell lines constitutively generated stimulatory activity for rat BM lymphoid precursor cells.However, after approximately 6 months of continuous activity, these cell lines lost the ability to spontaneously condition medium.This latter phenomenon, which subsequently has been observed with several other stromal cell lines, is not associated with obvious changes in proliferative activity or morphology of the stromal cells.Freshly harvested rat BM cells were incubated for 4 days in medium conditioned for 10 days with mouse BM-adherent cells (1x106 cells/ml).Sublethally irradiated rats were injected intravenously with 25 x 106 CM-treated cells for lymphocyte regeneration assay or x106 cells for CFU-S assay.Results in (A) represent the percentage (+S.D.) of donor-origin thymocytes (dark) and splenic B cells (lig t) 25 days postinjection, as determined by immunof- luorescence staining with RT-7 and IgK-1 alloreactive antibodies, respectively.Results in (B) represent the mean number (+ S.D.) of CFU-S per spleen 12 days postinjection.


DISCUSSION

Neither pre-B-cells nor their immediate progenitors are supported in culture inserts under Whitlock-  Witte culture conditions (Kierney and Dorshkind,   1987).However, cells that can give rise to pre-Bcells when transferred directly onto Whitlock-Witte   Mouse BM  by mouse BM stromal cell lines. 5x 105 culture-generated rat BM lyphoid cells were incubated for 4 days in 2 ml of the indicated CM.The mouse BM stromal cell lines (F12-11-B6, F1-5-B6) were produced as described in Methods.Results represent the number of total (solid) and TdT (hatched) BM lymphoid cells per well.Similar levels of activity were detectable in CM from the respective stromal cell lines for up to 6 months.Representative exper- iment (one of five).

feeder layers can survive in culture inserts under

Dexter-type culture conditions.Although the nature of these adherence-independent precursors is un- known, it is intriguing to speculate, given the results presented herein, that they may be related to the small number of TdT / lymphoid cells that we previously ave observed in standard Dexter cul- tures (Schrader et al., 1978).

Stem-cell factor, a regulatory mediator produced by stromal cells in both a soluble (Williams et al.,  1990; Zsebo et al., 1990) and membrane-bound (Flanagan and Leder, 1990; Huang et al., 1990)   form, appears to act synergistically with lineage- specific factors to stimulate the most immature members of a variety of hemopoietic cell lineages and may be involved in the adherence-dependent stage of pre-B-cell generation.However, membrane- bound SCF does not appear to play a major (or essential) role in our culture system, inasmuch as BM-adherent cells derived from Sl/Sl a mutant mice effectively support the generation of TdT / lymphoid cells in vitro (Medlock et al., 1987b).More- over, the failure to recover lymphostimulatory activity from the feeder layer by extraction with hypertonic saline suggests, but does not prove, that early lymphopoiesis is not enhanced by other me- diators that may be bound to cell membranes or ECM (Gordon et al., 1987; Roberts et al., 1988).

It therefore is of interest that decreasing the concentration of FBS in the culture medium mark- edly reduces the percentages of TdT / lymphoid cells in the adherent phase of the culture, but does not affect the total number of lymphoid cells that are generated (Medlock et al., 1993b).This observation permits two inferences: first, that a hyaluronatedependent adhesion system, regulated by serum (Matuoka et al., 1987) and similar to that described in Whitock-Witte cultures (Miyake et al., 1990), may be operational in our culture system; and, second, this adhesion system is not essential for optimal lymphopoiesis under conditions in which TdT lymphoid precursors can continue to contact the feeder layer.However, the lowered efficiency of early lymphopoiesis observed in culture inserts and in unconcentrated CM, and the continued adher- ence of most TdT-lymphoid cells under low serum conditions, leaves open the possibility that direct contact and/or adherence between at least a subset of lymphoid precursor cells and feeder-layer cells is important for optimal lymphopoiesis.Some investigators have found that direct contact between lymphoid cells and stromal cells induces the stromal cells to increase the level of production of cytokines such as IL-7 (Sudo et al., 1989).We therefore determined whether seeding of rat lymphoid precursor cells directly onto the mouse BM feeder layer could increase the number of TdT / lymphoid cells generated in culture inserts or CM by a similar mechanism.The observation that the number of TdT / cells in the culture inserts and CM was diminished rather than increased under these conditions suggests that the effective concentration of relevant soluble mediator(s) available to lymphoid precursors in culture inserts was not in- creased, possibly due to their preferential usage by the lymphoid cells in closest proximity to the feeder layer.This in turn suggests that the reduced effi- ciency of lymphopoiesis observed among rat BM cells placed in culture inserts is due, at least in part, to a suboptimal concentration gradient of soluble mediators.The dose-response effects on the num- bers of TdT / cells generated in CM further supports this notion.

As reported previously (Hayashi et al., 1984) and confirmed here, neither adherent ce l feeder layers from mouse thymus or spleen nor CM therefrom support early lymphopoiesis .inour culture system.This suggests that the lymphopoietic activity produced by mouse BM stromal cells in organ-specific.The failure of rat BM-adherent cell-feeder layers and CM therefrom to support lymphopoiesis in vitro does not contradict this thesis, inasmuch as rat BM-adherent cell-feeder layers are morphologically dissimilar to mouse BM feeder layers (Hayashi et al.,  1984; and unpublished observations).This observa- tion suggests that rat BM-derived microenvironmen- tal cells capable of supporting lymphopoiesis are not supported under the present culture conditions.It is of interest in considering the physiological relevance of our xenogeneic culture system that the mouse BM microenvironment is capable of supporting the de- velopment of rat lymphoid cells in vivo (Ildstad et  al., 1991, 1992) as well as in vitro.However, it should be cautioned that, in both instances, the regulation of rat lymphoid cell development by mouse BM stroma may be assisted by the presence of rat-origin microenvironmental cells engrafted along with the hemopoietic cells (Medlock et al.,  1987a).

A tentative in vivo model for B lymphopoiesis has been proposed in the rat, in which HIS24 / HIS50-c/-pro-B-cells (1.7% of total nucleated BM cells) give rise to HIS24 / HIS50 / c/-pre-pre-Bcells (5% of total nucleated cells), and then to HIS24 / HIS50 / c/ + pre-B-cells (20% of total nucle- ated cells) (Hermans, 1991).Results in the present study suggest that a similar developmental hierar- chy of rat early B-lineage cells exists in our culture system.Moreover, studies in which freshly har- vested or culture-generated HIS24 / HIS50-and HIS24 HIS50 / rat BM lymphoid cells were sepa- rated by flow cytometry and placed in culture indicate that the HIS50 / cells beget mostly HIS50 / cells and have a limited proliferative potential, whereas the HIS50-cells generate both HIS50- and HIS50 / cells and proliferate indefintely upon repeated passage (Goldschneider et al., 1989).It is of especial interest therefore that approximately 25% of the lymphoid cells recovered from CM and culture inserts in the present study had an ex- tremely primitive HIS24 / HIS50-TdT-c/-B- lineage phenotype, consistent with an even earlier, B220 / HSA-pre-pro-B-cells, stage of development described in the mouse (Hardy et al., 1991; Tong   et al., 1993).The results also suggest that the microenvironmental regulatory requirements for pre-pro-B-cells and pro-B-cell

for pre-
-cells.This is consistent with the recent observation (Funk and Witte, 1992) that c/ + pre-B-cells and TdT / pro-B-cells are located in anatomically istinct BM compartments in mice.

The distribution of TdT expression among the HIS50 / and HIS50-populations of culture- generated lymphoid

lls delinea
es four subsets of lymphoid precursors (Table 1) and permits further insights into the developmental pathway of early B-lineage cells.Thus, although all of the proliferat- ing cells in CM were HIS24 + HIS50-, approxi- mately half were also TdT + (Table3).When combined with the demonstration in previous stud- ies that the appearance of TdT-lymphoid cells precedes that of TdT + lymphoid cells in our culture system (Medlock et al., 1993b), as well as during ontogeny (Gregoire et al., 1979), the results strongly suggest that the least mature B-lineage cells in the nonadherent phase of our culture system (and the presumptive in vitro target of the soluble mediators in CM) are HIS24 / HIS50-TdT-pre-pro-B-cells and HIS24 / TdT + early pro-B-cells.The expression of TdT in the latter cells might then indicate preparation for D-JH Ig gene rearrangement (Desiderio et   al., 1984), the onset of which may be signified by the expressio of the HIS50 marker and cessation of cell proliferation at the intermediate pro-B-cells stage.It will be of considerable interest in this respect to determine whether the subset of HIS24 + HIS50 + TdT / lymphoid cells in our culture system corresponds to that which has undergone partial D-JI-I Ig gene rearrangement, whereas the subset of HIS24 / HIS50-TdT / lymphoid cells corresponds to that with germline D-JH configurations (Hunt et   al., 1988; Ehlich et al., 1993).If so, preparation for the synthesis of c/ presumably would follow the cessation of TdT expression in the HIS24 / HIS50 + cells (late pro-B-cells).Such a hypothetical model of the sequence of primitive B-lineage development in our culture system is presented in Fig. 9.

The most mature cells in the prior scheme of lymphopoiesis (HIS24 + HIS50 / TdT-late pro-Bcells) constitute up to 25% of the total lymphoid cells present in our culture system.Yet c pre-B- cells, their presumptive progeny, are produced only sporadically, even in the presence of exogenous IL-7  (unpublished observation).The reason for the fail- ure of these late pro-B-cells to synthesize c/ in vitro is being explored.However, many must be able to do so under the appropriate conditions, because culture-generated cells are approximately 25fold more efficient than are freshly harvested BM cells at producing sIg / B cells w.hen adoptively transferred in vivo (Goldschneider and McKenna, 1991).Fur- thermore, we have observed that a line of culture- generated intermediate pro-B-cells (HIS24 / HIS50- TdT/) expressed readily detectible c/ coincident with spontaneous leukemic transformation in vitro (unpublished observations).

Although only the lymphoid progenitor cells in the nonadherent compartment of the culture system were studied in the present experiments, it is likely that many of their counterparts in the adherent compartment are also responsive to the lymphostimulatory activity in CM and that cell contact is optional for most of these cells as well.Thus, we have observed that the lymphoid cells generated in culture inserts or in CM rapidly form foci of adher- ent lymphopoietic cells when seeded directly onto stromal cell-feeder layers; and that actively cycling lymphoid cells from the adherent and nonadherent phases of the culture system constitute phenotypi- cally indistinguish

le subpopulations (
ayashi et  al., 1984; Medlock et al., 1993b).Nonetheless, it is possible that the most primitive lymphoid precursors in these cultures require cell contact in addition to other signals for optimal stimulation, as suggested by the report of Hardy et al. (1991) for pre-pro-Bcells in the mouse.

To our knowledge, the present study is the first to demonstrate that populations of committed lymphoid stem/progenitor cells, devoid of detectible pluripotent hemopoietic stem cells (Hayashi et al.,  1984; and unpublished observations), can be generated for prolonged periods in culture inserts.Such cell populations appear to include prothymocytes as well as pro-B-cells, as demonstrated by in vivo adoptive transfer studies.Parallel experiments using medium co

itioned by stromal cell lines from mous
BM suggest that this early lymphopoietic activity is maintained by one or more stage-specific soluble mediators.The precise nature of this lymphostimulatory activity is presently being investi- gated.However, preliminary experiments indicate that the major factor is a novel high MW form of IL-7 that is bound, but not neutralized, by antibod- ies to IL-7.


MATERIALS AND METHODS


Animals

Male 4-6-week-old C57BL/6 strain mice, purchased from the National Cancer Institute (NCI) (Frederick, MD), were used as donors of BM adherent cells.Male 4-6-week-old Lewis strain rats, bred from Affinity-purified fluorescein isothiocyanate (FITC)conjugated goat IgG F(ab')2 anti-mouse and anti-rat IgG antibodies, and FITC-conjugated goat anti- mouse IgM (heavy-chain-specific) antibody were obtai

d from Kirkegaard and Perry Laboratories (Gai
hersburg, MD).Affinity-purified rabbit anti- body to calf thymus terminal deoxynucleotidyl transferase (TdT), and FITC-and tetramethyl- rhodamine isothiocyanate (TRITC)-conjugated goat anti-rabbit IgG were purchased from Supertechs (Bethesda, MD).Phycoerythrin (PE)-conjugated goat anti-mouse IgG was obtained from Caltag Labora- tories (San Francisco).


Immunofluorescence

Indirect immunofluorescence of cell-surface anti- gens was performed by incubating I x 106 freshly harvested or culture-generated cells with mouse or rat primary antibodies (10 1) and developing with appropriate FITC-or PE-conjugated goat anti-IgG or IgM antibodies (Hayashi, et al., 1984).To detect intranuclear TdT, cytocentrifuge-prepared cell smears were fixed in 4 C absolute methanol, stained with rabbit antibodies to TdT, and developed with FITC-or TRITC-conjugated a